Struct PromptLearningJob 

pub struct PromptLearningJob { /* private fields */ }

High-level prompt learning job orchestration.

Implementations

impl PromptLearningJob

pub fn from_dict( config: Value, api_key: Option<&str>, base_url: Option<&str>, container_worker_token: Option<String>, ) -> Result<PromptLearningJob, CoreError>

Create a job from a configuration dict.

Arguments

• config - Job configuration (algorithm, container_url, policy, etc.)
• api_key - Optional API key (uses env if not provided)
• base_url - Optional base URL (uses default if not provided)

Example

let job = PromptLearningJob::from_dict(
    serde_json::json!({
        "algorithm": "gepa",
        "container_url": "http://localhost:8000",
        "env_name": "default",
        "policy": { "model": "gpt-4o-mini", "provider": "openai" },
        "gepa": { "rollout_budget": 100 }
    }),
    None,
    None,
    None,
)?;

pub fn from_job_id( job_id: &str, api_key: Option<&str>, base_url: Option<&str>, ) -> Result<PromptLearningJob, CoreError>

Reconnect to an existing job by ID.

Arguments

• job_id - Existing job ID
• api_key - Optional API key
• base_url - Optional base URL

pub fn job_id(&self) -> Option<&str>

Get the job ID (if submitted).

pub fn tracker(&self) -> &ProgressTracker

Get the progress tracker.

pub async fn submit(&mut self) -> Result<String, CoreError>

Submit the job to the backend.

Returns the job ID on success.

pub async fn get_status(&self) -> Result<PromptLearningResult, CoreError>

Get the current job status.

pub async fn poll_until_complete( &self, timeout_secs: f64, interval_secs: f64, ) -> Result<PromptLearningResult, CoreError>

Poll until the job reaches a terminal state.

Arguments

• timeout_secs - Maximum time to wait
• interval_secs - Polling interval

pub async fn stream_until_complete<F>( &mut self, timeout_secs: f64, on_event: Option<F>, ) -> Result<PromptLearningResult, CoreError>

where F: FnMut(&ParsedEvent),

Stream events until completion with callback.

Arguments

• timeout_secs - Maximum time to wait
• on_event - Optional callback for each event

pub async fn cancel(&self, reason: Option<&str>) -> Result<(), CoreError>

Cancel a running job.

Arguments

• reason - Optional cancellation reason

pub async fn pause(&self, reason: Option<&str>) -> Result<(), CoreError>

Pause a running job.

Arguments

• reason - Optional pause reason

pub async fn resume(&self, reason: Option<&str>) -> Result<(), CoreError>

Resume a paused job.

Arguments

• reason - Optional resume reason

pub async fn get_results(&self) -> Result<PromptResults, CoreError>

Get detailed results including prompt extraction.

This fetches events to extract the best prompts.